Save Article Instructions

DARPA releases solicitation for artificial intelligence job to advance machine learning

ARLINGTON, Va., 4 April 2013. U.S. military researchers are approaching industry computer scientists for proposals for an artificial intelligence program to make machine learning capability much more accessible and effective for a wide variety of military weapons and information systems.

Scientists at the U.S. Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., released a formal solicitation this week (DARPA-BAA-13-31) for the Probabilistic Programming for Advancing Machine Learning (PPAML) program, which seeks rapid advancements in machine learning, which experts say is at the heart of modern approaches to artificial intelligence.

DARPA will brief industry on the PPAML next Wednesday in Arlington, Va. The goal of PPAML is to advance machine learning by using probabilistic programming to increase the number of people who can build machine learning applications; make machine learning experts more effective; and enable new applications that are impossible to conceive of using today’s technology.

Specifically, the PPAML program will try to make machine learning model code shorter; reduce development time; help build richer models; reduce the expertise necessary build machine learning applications; and support the construction of integrated models.

The DARPA PPAML program has four technical areas: domain experts; probabilistic programming; machine learning; and inference engine. Proposers may address any of the four areas.

The domain experts portion involves providing a machine-learning problem with a well-defined metric for evaluating the quality of solutions. Probabilistic programming involves building the front end of a probabilistic programming system that enables users from a range of skill levels to create useful machine learning applications from a variety of domains.

Machine learning, meanwhile, refers to performing basic research in machine learning relevant to the PPAML program. Inference engine, finally, involves building the back end of a probabilistic programming system that inputs models, queries, and prior data. DARPA officials say they plan to make several contract awards for each technical area except for the domain experts portion.

Related stories

-- DARPA takes another look at improving machine learning

-- DARPA pursues electronic warfare machine learning technology to automatically jam enemy adaptive communications

-- BAE Systems to develop electronic warfare machine learning to jam enemy adaptive communications automatically.

Machine learning assumes that enabling computers to learn can be more effective than programming them explicitly -- especially in military applications such as intelligence, surveillance, and reconnaissance (ISR); natural language processing (NLP); predictive analytics; cyber; and various scientific disciplines.

Example of machine learning include self-driving cars, image search and activity detection, object tracking, topic models, spam filters, recommender systems, predictive databases, and gene sequencing.

Although enabling computers to learn from their experiences can be valuable, building effective machine learning applications today requires Herculean efforts on the part of trained experts in machine learning, DARPA researchers point out.

Clearing this hurdle will research breakthroughs in several areas. For the front end of the system, the primary issues boil down to simplifying language while getting information efficiently; making the system usable by many people; devising profilers, debuggers, and model verification/checking tools to determine accuracy.

For the back end of the system, the key research challenge is improving system performance and predictability. To do this, DARPA researchers want to develop analyses that select the most appropriate solver for a particular query; improve the performance of existing solvers by blending in ideas from the compiler optimization community; compiling specific solvers for multi-core machines, graphics processing units (GPUs), cloud infrastructures, and custom hardware; developing new solvers; and developing an API for new solvers.

Research challenges in basic machine learning technology will include advancing the theory of probabilistic programming; discovering new efficient inference algorithms; discovering efficient representations; developing inference algorithms that work over streaming data; and developing techniques for assessing model fitness.

The PPAML industry briefings on 10 April are to familiarize participants with DARPA’s approaches to machine learning by using probabilistic programming; identify potential proposers; and promote teaming. Companies that want to attend the briefings should register for the PPAML briefings no later than tomorrow, 5 April, online at For questions or comments email DARPA at

Companies interested in submitting bids for the PPAML program should respond to DARPA no later than 16 May 2013. For questions or concerns email Kathleen Fisher, the DARPA PPAML program manager, at

More information about the PPAML solicitation is online at

To access this Article, go to: